Alternative to domain wall fermions

SummaryThis paper reviews the most popular methods which are used in lattice QCD to compute the determinant of the lattice Dirac operator: Gaussian integral representation and noisy methods. Both of them lead naturally to matrix function problems. We review the most recent development in Krylov subspace evaluation of matrix functions. The second part of the paper reviews the formal relationship and algebraic structure of domain wall and overlap fermions. We review the multigrid algorithm to invert the overlap operator. It is described here as a preconditioned Jacobi iteration where the preconditioner is the Schur complement of a certain block of the truncated overlap matrix.

We briefly review1 the overlap formalism for chiral gauge theories, the overlap Dirac operator for massless fermions and its connection to domain wall fermions. We describe properties of the overlap Dirac operator, and methods to implement it numerically. Finally, we give some examples of quenched calculations of chiral symmetry breaking and topology with overlap fermions.

We present renormalization constants of overlap quark bilinear operators on 2+1-flavor domain wall fermion configurations.Both overlap and domain wall fermions have chiral symmetry on the lattice.The scale independent renormalization constant for the local axial vector current is computed using a Ward Identity.The renormalization constants for the scalar, pseudoscalar and vector current are calculated in the RI-MOM scheme.Results in the MS scheme are obtained by using perturbative conversion ratios.The analysis uses in total six ensembles with lattice sizes 24 3 × 64 and 32 3 × 64.

We discuss the symmetries of quenched QCD with Wilson fermions, starting from its lagrangian formulation, taking into account the constraints needed for convergence of the ghost-quark functional integral. We construct the corresponding chiral effective lagrangian, including terms linear and quadratic in the lattice spacing. This allows us to study the phase structure of the quenched theory, and compare it to that in the unquenched theory. In particular we study whether there may be an Aoki phase (with parity and flavor spontaneously broken) or a first order transition line (with no symmetry breaking but meson masses proportional to the lattice spacing), which are the two possibilities in the unquenched theory. The presence of such phase structure, and the concomitant long-range correlations, has important implications for numerical studies using both quenched and dynamical overlap and domain-wall fermions. We argue that the phase structure is qualitatively the same as in the unquenched theory, with the choice between the two possibilities depending on the sign of a parameter in the low-energy effective theory.

Within the bottom-up holographic approach to QCD, the highly excited hadrons are identified with the bulk normal modes in the fifth “holographic” dimension. We show that additional states in the same mass range can appear also from taking into consideration the 5D fields dual to higher dimensional QCD operators. The possible effects of these operators have not been taken into account in virtually any phenomenological applications. Using the scalar case as the simplest example, we demonstrate that the additional higher dimensional operators lead to a large degeneracy of highly excited states in the soft wall holographic model, and in the hard wall holographic model, they result in a proliferation of excited states. The considered model can be viewed as the first analytical toy model predicting a one-to-one mapping of the excited meson states to definite QCD operators, to which they prefer to couple.

A class of d-dimensional Dirac operators with a variable mass is introduced (), which includes, as a special case, the 3-dimensional Dirac operator describing the chiral quark soliton model in nuclear physics, and some aspects of it are investigated.

In this thesis we introduce the S-theory. We apply the S-theory to variational problems of immersed Riemann surfaces, Elasticity theory and Spin theory. 
\nFrom classical mechanics it is known that geometric problems can be substantially simplified by symmetry considerations. The Noether theorem describes in this case how to find conservation laws from the invariance property of the observed energy. At first, the motivation was to give the space of immersions M = {f : M → R3}, for an oriented Riemannian manifold M of dimension less than or equal to three, an Euclidean invariant description. For geometric energies, i.e. energies which are invariant under Euclidean transformations, one would be able to derive conservation laws for the respective variational problems. It is a well known fact that immersed space curves are uniquely determined by their curvature functions (κ1,κ2,τ) up to Euclidean transformations. In the case of immersed surfaces and immersed 3-manifolds the situation becomes more complicated and leads to a non conformal deformation theory (S-Theorie). 
\nThe S-Theory builds the foundations of this thesis. Starting from a Riemannian manifold M and a reference metric ⟨ , ⟩ one can model any other Riemmanian metric g through a positive definite and self adjoint operator S via g = ⟨S,S⟩. 
\nThe operator S is an isometry, i.e. g = S∗⟨,⟩. From this point of view we compute the Levi- Civita connection with respect to g out of Levi-Civita connection of the reference metric and the operator S. 
\nFurther we introduce Spin bundles over two and three dimensional oriented Riemannian manifolds. Thereby Spin bundles, in contrast to most of the literature, have additionally a quaternionic structure. Many formulas become clearer and more accessible. We show that any Spin bundle Σ over (M, ⟨ , ⟩) has a unique Spin connection and compute via the S-Theory the deformed Spin connection with respect to the metric g. Eventually we consider the induced Spin bundle of an immersion f : M → R3 and compute the corresponding Spin connection and find a new interpretation of the Gauss-Codazzi equation. We introduce the Dirac operator and compute its deformation with respect to the metric g. 
\nWe are now able to give the space of immersed surfaces M = {f : M → R3} the desired geometric description. We compute the normal space of M and formulate the Noether theorem. As an application we compute conservation laws for the Willmore functional. 
\nFinally we introduce the intrinsic version of the famous distance-squared energy on a n-dimensional Riemannian manifold. We introduce the corresponding stress tensor and show that it’s closeness is a characterization for critical points of the energy. Thereby the S-theory provides a fundamental concept for the understanding of the stress tensor. In the case of a Riemann surface, the S-theory provides a canonical harmonic rotation 1-form for critical points of the distance-squared energy.

Random matrix theory and the spectra of overlap fermions

We discuss how to construct anomaly-free chiral gauge theories on the lattice with exact gauge invariance in the framework of domain wall fermion. Chiral gauge coupling is realized by introducing a five-dimensional gauge field which interpolates between two different four-dimensional gauge fields at boundaries. The five-dimensional dependence is compensated by a local and gauge-invariant counter term. The cohomology problem to obtain the counter term is formulated in 5+1 dimensional space, using the Chern-Simons current induced from the five-dimensional Wilson fermion. We clarify the connection to the invariant construction based on the Ginsparg-Wilson relation using overlap Dirac operator. Formula for the measure and the effective action of Weyl fermions are obtained in terms of five-dimensional lattice quantities.

Spectral analysis of dissipative Dirac operators with general boundary conditions

We present an exact pseudofermion action for hybrid Monte Carlo simulation (HMC) of one-flavor domain-wall fermion (DWF), with the effective 4-dimensional Dirac operator equal to the optimal rational approximation of the overlap-Dirac operator with kernel H=cHw(1+dγ5Hw)−1, where c and d are constants. Using this exact pseudofermion action, we perform HMC of one-flavor QCD, and compare its characteristics with the widely used rational hybrid Monte Carlo algorithm (RHMC). Moreover, to demonstrate the practicality of the exact one-flavor algorithm (EOFA), we perform the first dynamical simulation of the (1+1)-flavors QCD with DWF.

The feasibility of using lattice chiral fermions which are free of O(a) errors for both the heavy and light quarks is examined. The fact that the effective quark propagators in these fermions have the same form as that in the continuum with the quark mass being only an additive parameter to a chirally symmetric anti-Hermitian Dirac operator is highlighted. This implies that there is no distinction between the heavy and light quarks and no mass dependent tuning of the action or operators as long as the discretization error O(m2a2) is negligible. Using the overlap fermion, we find that the O(m2a2) (and O(ma2)) errors in the dispersion relations of the pseudoscalar and vector mesons and the renormalization of the axial-vector current and scalar density are small. This suggests that the applicable range of ma may be extended to ~0.56 with only 5% error, which is a factor of ~2.4 larger than the corresponding range of the improved Wilson action. We show that the generalized Gell–Mann–Oakes–Renner relation with unequal masses can be utilized to determine the finite ma corrections in the renormalization of the matrix elements for the heavy-light decay constants and semileptonic decay constants of the B/D meson.

The inclusion of higher-dimensional gauge invariant operators induces new Lorentz structures in Higgs couplings with electroweak gauge boson pairs. This in principle affects the kinematics of Higgs production and decay, thereby modifying the efficiencies of the experimental cuts compared to what simulations based on the standard model interactions yield. Taking some sample cases, we perform a rigorous analysis of how the efficiencies differ for various strengths of the additional operator vis-a-vis the standard model interactions, scanning over the values of both of them. While the response to cuts can be markedly different in some regions, we find that the sensitivity to new operator structures is relatively limited, so long as we remain confined to the 2-sigma regions around the best fit signal strengths measured at the Large Hadron Collider. We also show modifications to certain kinematical distributions including the new operators in the diphoton final state.

We study the renormalization of the DeltaS=1 effective weak Hamiltonian with\noverlap fermions. The mixing coefficients among dimension-six operators are\ncomputed at one loop in perturbation theory. As a consequence of the chiral\nsymmetry at finite lattice spacing and of the GIM mechanism, which turns out to\nbe quadratic in the masses, the K-->pi pi and K -->pi matrix elements relevant\nfor the Delta I =1/2 rule can be computed without any power subtractions. The\nanalogous amplitudes for e'/e require one divergent subtraction only, which can\nbe performed non-perturbatively using K-->0 matrix elements.\n

We formulate a theorem saying that the Dirac operator corresponding to a one-electron atomic ion is essentially self-adjoint on the usual domain, provided that the nuclear charge Z is less than 118. Furthermore, for such nuclear charges the domains of the closure of the free particle and total Dirac operators are equal. In the present part I of this paper we prove this theorem for the part of the operator over each of the usual reducing subspaces.